Urban Growth and Violence: Will the Future Resemble the Past?

Summary

Many social, economic, and political problems have accompanied urban growth in the developing world. Will further growth result in violent behavior as expectations of economic improvement and social mobility are dashed? Past theoretical and empirical research on the links between urban growth and violence showed the relationship to be weak to nonexistent. But these past findings may not apply to present and future urban environments. This paper suggests why the future may differ from the past. In particular, it points to factors that may interact with urban growth in the future to lead to a more violent urban experience. These factors include periodic economic crisis, the reduction of state capacity to cope with political challenges, grassroots demands for democratization, and a gradual fading of the rural experience as a basis for evaluating relative economic standing and opportunity.

 

Full Article

Does rapid urban growth contribute to urban violence? The story is more complicated than commonly thought. This paper argues that, by itself, urban growth is quite benign. However, in interaction with other factors, such as economic crises and a weak state, urban growth appears much more likely to contribute to violence.

Urban growth is continuing, in some places rapidly, and urban size is becoming truly astonishing in some parts of the developing world. Between 1950 and 1990, there was a fivefold increase, to 1.5 billion, in the number of urban residents in developing countries; about 37 percent of the population of the Third World now lives in cities. By 2025, the United Nations projects a further tripling of the total to 4.4 billion, at which point nearly two-thirds of the citizens of the developing world will live in cities.¹

In the aggregate, cities in the developing world are growing by an estimated 160,000 persons per day. Mexico City, which had a population of 3.1 million in 1950, is projected to reach over 25 million by the end of the 1990s; estimates for São Paulo suggest a population of 22 million by the year 2000 — almost 10 times the 1950 total. While it took New York City almost 150 years to grow by 8 million people, Mexico City and São Paulo will add this number in less than 15 years. In 1970, there were only 3 megacities in Asia with more than 8 million inhabitants; by the year 2000, there will be 17. The number of cities with at least 1 million inhabitants has gone from 31 in 1950 to 180 in the early 1990s and is expected to rise to more than 300 by the end of the century.

Large and dynamic cities offer many benefits to developing societies; social scientists have long recognized that cities provide exceptional opportunities for entrepreneurship, creativity, and the generation of wealth. But a host of intractable problems often accompanies rapid urban growth. Kasarda and Parnell note that these problems include:

. . . high rates of unemployment and underemployment as urban labor markets are unable to absorb the expanding number of job seekers, soaring urban poverty, insufficient shelter, inadequate sanitation, inadequate or contaminated water supplies, serious air pollution and other forms of environmental degradation, congested streets, overloaded public transportation systems, and municipal budget crises.²

What effects will urban growth and its consequences have on civil stability, in particular on the incidence of mass violence? This paper addresses this question. First, we briefly review the theory and evidence on the nature of urban violence and the sources of urban growth. We then consider the insights and shortcomings of past research on links between urban growth and violence. This early work, we argue, suffers from oversimplification and narrow focus. Some recent anecdotal evidence suggests that urban growth may contribute to violence as an interactive variable by amplifying the effects of other forces that are potential causes of urban conflict. We identify several such forces that may interact with urban growth to produce a more violent urban experience than existed in the past. In the paper’s conclusion, we discuss future research directions and identify some potential hypotheses that merit close study.

THE NATURE AND INCIDENCE OF URBAN VIOLENCE

Urban violence refers to the destruction of persons and property within an urban context.³ Any treatment of the causes of urban violence must ask who are its perpetrators, what are their purposes, and who or what are their targets.

Categories of Urban Violence

We can identify at least three broad categories of urban violence that are relevant to any discussion of the origins of urban conflict and its effects:

1. political violence, involving both violence directed against the state and violence by the state against challengers;
2. communal and ethnic violence; and
3. criminal and anomic violence.

Researchers concerned with the study of civil violence — that is, violence within countries — have mainly focused on the first category; and almost all of this attention has focused on collective violence directed against the state. Such violence is often the product of mass unrest and dissatisfaction with state performance, and it includes riot, insurgency, rebellion, revolution, and civil war. In response, the state may itself resort to violence to address these overt challenges to its authority. Yet states may also use strong-arm tactics, such as intimidation, torture, and assassination, to prevent challenges from ever occurring.

The second category is urban violence involving rival ethnic, racial, or religious groups. This appears to be an evermore conspicuous form of violence in today’s world. Here, the protagonists are generally private parties,⁴ yet the issues in contention and the violence itself may have great political consequences for the state and society in general. Such rivalries often involve perceived disparities in access to political and economic opportunities. Racial, ethnic, or religious identities serve as rallying points for political mobilization to address these disparities.

Least overtly political are those acts falling under the rubric of criminal and anomic violence. Wanton acts of destruction, armed robbery, assault, murder, and racketeering by individuals and groups fall into this category. The sources of criminal activity vary, but such actions are not necessarily devoid of political significance. Theorists from Emile Durkheim to Chalmers Johnson argue that an erosion of society’s moral unity is a key precursor to civil violence.⁵ To the extent that criminal or anomic violence reflects alienation from society, or a calculation that the potential gains of ignoring society’s rules exceed the costs of doing so, it indicates a breakdown in the moral and coercive authority of society as a whole.

Evidence on Rates and Trends of Violence

The developing world offers many examples of the urban violence described above. In terms of political violence directed against the state, during the mid-1970s, austerity measures adopted by debt-ridden countries sparked a wave of urban protest worldwide. Between 1976 and 1992, over 146 separate incidents of strikes, riots, and demonstrations took place — principally in Latin America.⁶ Some modern revolutions have also had an urban base. The Bolivian revolution in 1952 found its roots and support in organized labor and the disaffected middle class, with rural elements uninvolved until the government’s overthrow.⁷ In Iran in the late 1970s, student street demonstrations and labor protests in major cities were rallying points for the overthrow of the shah.⁸ And in Nicaragua in 1979, urban insurrection by workers and disaffected youth was crucial to the success of an organized guerrilla force against the Somoza regime.⁹

Ethnic and communal strife is widespread in many Third World cities. In Karachi, Pakistan, Sunni Muslim militants and Shiite extremists attack each other’s buses and mosques. Last year, this struggle claimed the lives of over 40 police officers and 70 political activists.¹⁰ Many Indian experts believe that cities are fertile breeding grounds for communal conflict; incidents of violence and brutality are far more common than in villages. Communal issues often become vents for the pent up anger and frustration produced by the high tension of urban life. In 1992, the demolition of the Babri Masid Mosque caused an explosion of Hindu-Muslim violence in many Indian cities. Of the 1,500 who died, almost 95 percent perished in urban areas.¹¹ Worst hit were the cities of Ahmedabad and Bombay, with gang rapes, murders, and acts of arson continuing months after the demolition. Similar incidents also occurred in Surat, Calcutta, Bhopal, and Bangalore.

Indian statistics, shown in Table 1, indicate that since the 1950s there has been a steady increase in the frequency of communal incidents and in the number of persons killed and injured as a result.¹² The frequency has increased rapidly in rural areas, but the majority of incidents remain urban.¹³ Moreover, the rate of increase has been faster than the rate of either rural or urban population growth, which means that the per capita incidence of communal violence has sharply increased.

Criminal and anomic violence often accompanies communal strife. In Karachi, rates have jumped sharply in recent years. In 1991, police reported 466 murders, 802 attempted murders, 421 cases of rioting (many against civic agencies), 103 rapes, and 140 kidnappings for ransom; in addition, an estimated 50 percent of all crimes go unreported. During communal incidents, the young of the idle rich sometimes turn to crime for excitement.¹⁴ In India, communal riots have offered pretexts for personally motivated violence. In one incident, property sharks in Calcutta took advantage of communal disorder to instigate the destruction of a lower-middle-class Hindu colony so that a shopping complex could be built on the land at a later date.¹⁵ Similar private exploitation of interethnic violence accompanied the clashes between rival black ethnic groups in South Africa during the late 1980s and early 1990s.¹⁶

Table 1: Communal Incidents in India, 1954-85

Year	No. of Incidents	Persons Killed	Persons Injured
1954	84	34	512
1955	75	24	457
1956	82	35	575
1957	58	12	316
1958	40	7	369
1959	42	41	1344
1960	26	14	262
1961	92	108	593
1962	60	43	348
1963	61	26	489
1964	1070	1919	2053
1965	173	34	758
1966	144	45	467
1967	198	251	880
1968	346	133	1309
1969	519	673	2702
1970	521	298	1607
1971	321	103	1263
1972	240	69	1056
1973	242	72	1318
1974	248	87	1123
1975	205	33	890
1976	169	39	794
1977	188	36	1122
1978	230	110	1853
1979	304	261	2379
1980	421	372	2691
1981	319	196	2613
1982	470	238	3025
1983	500	1143	3652
1984	476	445	4836
1985	525	328	3665
Source: P. R. Rajgopal, Communal Violence in India (New Delhi: Uppal Publishing House, 1987), pp. 16–17.

Other cities such as Bogota, Colombia; Lagos, Nigeria; Dacca, Bangladesh; and San Jose, Costa Rica show the serious threat that criminal violence can pose on its own. All these cities have seen crime increase in recent years.¹⁷ Rio de Janeiro, Brazil’s second-largest city, is now plagued by slums, poverty, and urban violence. Today, it leads the country in negative urban indicators — with the largest number of slum dwellers (1 million), the highest murder rate (1 of 700 residents per annum), and the highest kidnapping rate (4 per week).¹⁸ In 1989, the city’s homicide rate was three times higher than New York’s, and the rate of urban violence continues to rise. Almost one-quarter of all homicides occur among people between the ages of 10 and 19. Amid the squalor and decay, organized crime has gained a foothold. Drug lords have established retail centers in the city’s shantytowns (favelas), while children serve as sentries and distributors of the illegal product. Male teenagers and young adults are recruited into death squads in the war against rivals and authorities.¹⁹

All such violence risks invoking the coercive power of the state. The official response follows a common sequence: police are first strengthened to restore law and order, then the national guard or special security forces are deployed, and, ultimately, the army is sent in if the preceding measures have not succeeded. The state often co-opts the media and uses them for counterpropaganda. But even when these tactics are successful, frequent use of state repression and violence to restore order indicates a decline in state legitimacy and capacity to govern.

THE NATURE AND CAUSES OF URBAN GROWTH

Three major factors account for urban population growth: natural growth, net migration, and reclassification.²⁰ Of these, the first two contribute the most.

Types and Causes of Urban Growth

Natural growth results from an excess of births over deaths within a city; this is growth caused by the natural reproduction of the city’s residents. Net migration produces urban growth when migration into the city exceeds migration out of the city. Migrants into a city usually share the same nationality as their urban-born counterparts and originate from the country’s rural areas. Many things attract these people to the city: most importantly, they may need to escape a rural environment increasingly incapable of sustaining them, and they may be attracted by an urban environment that seems to offer a better standard of living.

The two motivations are often related but are nevertheless distinct. In the first instance, the rural environment acts as a “push” factor on migrants. Dire rural conditions are sometimes attributed to rapid population growth — excess labor supply in the countryside aggravates rural poverty, providing a “push” to the cities.²¹ Recent work suggests that rural out-migration can also stem from generally low economic growth rates, shifts in agricultural productivity, and inequalities in rural land tenure.²² Factors such as climate, land availability, level of agricultural technology, demand for agricultural products and the availability of credit, fertilizers, and technical assistance are thus key in determining the degree to which the rural economy can absorb a growing rural labor force.²³

Cities also “pull” migrants in their direction by providing better employment and income opportunities than the countryside. Large urban areas are often the major engines of wealth generation in developing economies, providing dense and synergistic concentrations of capital, talent, and entrepreneurial opportunity. Industries tend to locate themselves in urban areas.²⁴ Markets are large, which means industries can take advantage of economies of scale, and cities provide convenient access to domestic and international markets through diverse transportation systems.²⁵ Industries also benefit from access to highly specialized financial, legal, and technical services. The marginal costs of increments to urban infrastructure are also lower in large cities, which encourages governments to provide basic services.

In recent years, scholars have recognized that much rural-urban migration in developing countries is “circular.” Some migrants move to the city for relatively short periods and then return to the country. Often this circular migration is prompted by changing availability of seasonal employment in the countryside and the city; it also arises from migrants’ desires to remain in contact with their home communities. Analysts note that circular or transient migration makes it hard to estimate the relative contribution to urban growth of movement from rural areas. Circular migration by a single member of a rural family is often the first step in the movement of the family, and eventually of whole communities, into the city.²⁶

Obviously, the relative contributions to total urban growth of natural growth and permanent migration vary in different areas of the developing world. Yet migration generally contributes more at early stages of economic development, with natural growth predominating during intermediate stages.²⁷ During later stages, rates of natural growth tend to drop and net migration once again emerges as the dominant factor in the growth of urban population.²⁸

Evidence on Urban Growth and Its Causes

Urban growth rates in much of the developing world have declined slightly from the very high rates that prevailed in the 1960s and 1970s. Table 2 and the Appendix give recent data on trends by region and country.

Table 2: Urban Growth Rate by Region

Region	1965-80	1980-90
Sub-Saharan Africa	5.8	5.9
East Asia	4.3	3.3
China	2.3	13.5
South Asia	3.9	3.9
Middle East/North Africa	4.6	4.4
Latin America	3.9	3.0
Source: Alan Gilbert, “Third World Cities: The Changing Development System,” Urban Studies, Vol. 3, Nos. 4–5 (May 1993), p. 722.

Data from a 1990 megacity survey by the International Labor Organization and from the UN Department of International Economic and Social Affairs indicate that the contribution of rural-urban migration to current urban growth in developing countries varies widely across countries (Tables 3 and 4). Although migration’s contribution is below 50 percent for most large cities in the developing world, it still contributes a large share. Migration usually contributes more to urban growth than does natural growth in the early and late stages of a society’s economic development. Therefore, since many Third World countries are still in an intermediate stage of development, we can expect higher relative contributions from migration in the future.²⁹

Table 3: Net Migration Contribution to Urban Growth, Selected Cities, 1960–85

City	1960–70	1970–80	1980–85
Bombay	48.8	47.7	–
Cairo	23.1 (1960-76)	–	–
Lagos	–	61.1	–
Manila	–	36.2 (1975–80)	–
Mexico City	–	45.8	28.4
Rio de Janeiro	26.5	39.5	–
Seoul	76.2	60.9	54.1
Shanghai	–	24.9	37.6 (1981–88)
Source: International Labor Organization megacity survey (1990); adapted from A.S. Oberai, Population Growth, Employment and Poverty in Third World Mega-Cities (New York: St. Martin’s Press, 1993), p. 27.

Table 4: Net Migration Contribution to Urban Growth, Selected Countries, 1975–90

Country	Net Urban In-Migrant Rate (Thousands)	Urban Population (Thousands)		Average Annual In-Migrants (Thousands)	Migrant Share of Urban Growth (Percent)
	1970-90	1975	1990
Africa
Kenya	46	1775	5923	177	64.17
Senegal	27	1631	2831	60	75.16
Tanzania	79	1602	8971	417	84.96
Tunisia	25	2673	4439	91	76.84
Latin America
Brazil	21	66793	115674	1934	59.35
Colombia	9	11899	22371	149	21.35
Costa Rica	23	830	1617	28	52.59
Cuba	15	5993	7736	104	89.50
Ecuador	23	2984	6136	105	49.91
Guatemala	12	2231	3861	37	34.34
Honduras	26	996	2240	42	50.33
Panama	9	858	1324	10	30.73
Peru	21	9313	15681	262	61.82
Puerto Rico (US)	15	1879	2742	35	61.04
Uruguay	2	2348	2673	5	23.17
Asia
Bangladesh	40	6985	15759	455	77.77
Fiji	11	212	329	3	37.97
Indonesia	27	26259	51975	1056	61.61
Iran	22	15240	31066	505	47.84
Iraq	18	6764	14034	187	38.62
Philippines	18	15136	26432	374	49.68
South Korea	29	16947	31397	701	72.77
Sri Lanka	-1	2998	3677	-3	-7.37
Thailand	31	6283	12609	293	69.43
Source: Sally E. Findley, “The Third World City: Development Policy and Issues,” in John D. Kasarda and Allan M. Parnell, eds., Third World Cities: Problems, Policies and Prospects(London: Sage, 1993), p. 15.

We may be seeing disturbing changes in the causes of rural-urban migration. Some experts argue that, historically, pull factors have predominated: migrants have viewed the urban environment as providing better employment and income opportunities. But in some regions of the world, especially in Africa and South Asia, push factors seem to be increasingly powerful.³⁰ Particularly important are changes in agricultural productivity and land-tenure systems that give rise to rural poverty, severe social inequalities, and landlessness. Scarcities of renewable resources, including cropland, forests, and water, also drive migration.³¹

The extent to which economic expectations are satisfied in the urban environment is open to question, especially in Africa, which continues to experience economic recession. Nonetheless, the urban influx in Africa remains large. And in the case of China, while reported growth rates are somewhat suspect,³² large-scale migration from the countryside is resulting in a glut of unemployed labor, which raises doubts about the absorptive capacities of even fast-growing urban economies.³³ Rural-urban migrants tend to be relatively young, which accentuates the youth bulge in urban populations in poor countries (Table 5). In Cairo, for instance, more than 40 percent of the population is under 15 years of age. This age distribution contributes to natural growth in urban populations, because a younger population has more children and fewer deaths compared with an older population, even though individual mothers in urban areas usually have lower fertility rates than their rural counterparts.³⁴ Furthermore, young urban populations, especially unemployed young men, are easier to mobilize politically, and young populations generate enormous demands for the provision of social resources such as education and jobs.

Table 5: Proportion of Population in the 0–19 Age Group Cohort, Selected Cities

Developed Countries	Year	Percent	Developing Countries	Year	Percent
Amsterdam	1980	22.1	Bangkok	1981	44.1
Birmingham	1980	29.9	Bombay	1981	41.5
Frankfurt	1981	24.8	Cairo	1986	44.4
London	1981	27.6	Delhi	1980	48.9
Los Angeles	1980	28.8	Jakarta	1981	52.9
Madrid	1980	33.5	Lagos	1985	45.7
Montreal	1980	23.3	Mexico City	1985	47.0
New York	1980	28.1	Rio de Janeiro	1980	36.5
Paris	1982	18.7	São Paulo	1980	40.0
Rome	1981	29.6	Seoul	1980	42.5
Tokyo	1981	28.2	Shanghai	1988	24.3
Source: A. S. Oberai, Population Growth, Employment and Poverty in Third-World Mega-Cities (New York: St. Martin’s Press, 1993), p. 28.

Hypotheses and Evidence on the Links Between Urban Growth and Violence

Policymakers and scholars have long expressed concerns about rapid urban growth, the ability of societies to adapt to such growth, and the possibility that it might contribute to political instability and urban violence.

Three Hypotheses

Arguments about urban growth and violence date to the 1960s and to concerns that rising political demands would inevitably accompany the concentration of population in Third World cities. Political and social mobilization caused by urban growth could outstrip the institutional capacity of the state and society.³⁵ Some scholars also suggested that rapid rural-urban migration would produce frustration of migrants’ expectations of economic improvement and social mobility or could cause processes of personal and social disorientation and alienation.³⁶ The outcome would be rising discontent and ultimately violent behavior – primarily directed against the state. The focus was therefore principally on political rather than on communal/ethnic or criminal violence.

Scholars disputed whether such violence would quickly follow on the heels of migration or erupt some years after migrants had been absorbed into the urban environment.³⁷ Early arguments merely warned of “disruptive migrants,” uprooted from rural roots, isolated in the city, and prone to violence and extremism as a result of increasing disillusionment with urban life. Later arguments suggested that violence was more likely after migrants had become more firmly established. Only after they shed their rural outlook and sense of strangeness and diffidence would their inability to find decent jobs and housing translate into the strong sense of deprivation and frustration required to prompt violent political action. In short, a socialization period was important for transforming migrants into “radicalized marginals.”

Nonetheless, the psychological processes that scholars suggested would move migrants to violent action were largely the same for both arguments. According to Wayne Cornelius, three propositions were central:

1. Rural-urban migration breeds economic frustration among migrant populations. The rapid influx of migrants cannot be accommodated by public or private sectors. Mobility expectations are thwarted, and the proximity of conspicuous consumption by elites raises migrant awareness of their marginal role in society. Migrants therefore experience rising relative deprivation,³⁸ which increases their propensity to engage in radical political activity (Figure 1).³⁹

2. Migrants have problems adjusting socially and psychologically to the urban environment (Figure 2). Cultural conflicts and the disruption of past living habits and customs cause personal identity crises, which increase the chances of primary group breakdown. Traditional social controls on deviant behavior are thus weakened. In addition, as migrants seek entry into new protective groups, they enter a phase in which they are susceptible to recruitment into extremist political movements.⁴⁰

3. Rural-urban migration, increased political awareness, and mobilization of radical opposition go hand in hand (Figure 3). The urban environment allows high levels of social communication and produces intense competition among various interest groups. Organized political activity is conspicuous, which helps politicize migrants and encourages mass involvement in political action.⁴¹ These factors translate into migrant support for opposition political parties and aggressive protest movements.

The above three processes could presumably occur simultaneously – with the first and second causing certain types of grievance (i.e., deprivation, feelings of alienation, anomie, and rootlessness) among migrants, and the third expanding the opportunities for those grievances to be articulated and acted upon by the people experiencing them (i.e., by facilitating social communication and comparisons with others and by altering the balance of power in society as challenger groups become stronger relative to the state).

Such arguments complement a range of work on the origins of violent conflict. Political, social, and economic disparities between classes or ethnic groups have long been thought to generate tension. They are often discussed in theoretical literature on relative deprivation and on the “moral economy” between rulers and the ruled.⁴² Both of these perspectives assume that people and groups feel aggrieved when their expectations about rightful entitlements are not met; and both view such expectations, and, in turn, the degree of grievance generated by their lack of realization, as derived from certain subjective standards of equity, fairness, and justice.

By shrinking the distance between economic classes and ethnic groups and by generating demands that cities cannot meet, rapid urban growth might make inequalities and injustices more obvious. It might aid social comparisons that generate feelings of deprivation and perceptions that the moral economy between powerful and weak groups has been violated. At the same time, urban growth can open up opportunities for political mobilization, organization, and action by making communication easier and by promoting more vigorous social interaction among aggrieved people and groups.

Survey of Current Evidence

Early research on links between urban growth and violence yielded little empirical support. For instance, fluctuations of collective political violence in the United States, Britain, France, and Mexico did not correspond to the pace of urbanization.⁴³ In the United States, the 1967 Presidential Commission on Civil Disorders found that migrants were less involved in violence than were people raised in cities and that there was not much evidence of a positive relationship between urban growth and crime. A correlation between population growth and property crime was found in some cases, yet it was not high and may have been better associated with city size.⁴⁴

There are a number of explanations of these findings. Migrants from rural areas might be ignorant of structural obstacles to their mobility in cities, and therefore might not develop strong feelings of deprivation and injustice. The conservatism and acquiescence common among migrants could stem from rural values that promote deference and political passivity. Recent migrants also tend to be preoccupied with acquiring the basic necessities of life, which generally means that they must work through existing institutions rather than challenge them.⁴⁵ And the nature of social organization in urban slums may create feelings of distrust among subgroups within migrant communities.⁴⁶ The communities are usually organized around clearly defined urban territories and neighborhoods; this very localized and insular culture does not lend itself to mobilization based on broader groupings and more universal ideas.

Other forces could defuse and constrain protest as well. Most obvious is the power of governments to address grievances, co-opt dissent via patronage, and use repression against challengers. Less obvious, but nonetheless significant, is the ability of certain cultures to resist the deterioration of family and kinship ties expected as a result of rural-urban migration. In addition, divisions along ethnic, religious, and caste lines could prevent the emergence of effective mass movements – by obscuring economic stratification and defusing class conflict.

Yet early studies also revealed evidence that flatly contradicted the above arguments. Instead of feeling deprived, many migrants were satisfied with urban life in comparison with the rural existence they had left. In fact, many viewed conditions in the city as much better than in the country. Migrants seemed to find jobs quickly, often securing employment prior to their arrival. Data indicated relative improvements in migrant family income levels.⁴⁷ In addition, migrant occupation patterns were found to differ little from those of the urban born.⁴⁸ Tested less was the hypothesis that a period of socialization was required before migrant radicalization could occur; it was difficult to get data tracking migrants over time.

Overall, the early research clearly showed that simple arguments about the links between rural-urban migration and violence were misguided. Although these arguments acknowledged that rapid urban growth could overburden state institutions and generate grievances among the urban populace, they took little account of the many ways differing political systems, societies, and cultures could cope with migration to urban areas. Nor did these arguments adequately distinguish among the types of violence that might occur as a result of rural-urban migration; they generally focused on mass political violence to the exclusion of other forms of urban strife. Moreover, the early arguments said virtually nothing about the violence-inducing potential of urban growth in general or about the international and domestic contexts in which states and their urban centers were embedded.

It is also possible that the negative findings of early research resulted from the choice of cases. In particular, most early research studied rural-urban migration during periods of relative political and economic stability (if not growth) in the societies affected. States and institutions may therefore have been able to respond more effectively to migrant demands, and these demands may have been lower than would have been the case in more difficult economic times. In fact, after noting a weak relationship between rural-urban migration and violence, early researchers often argued that continued political stability depended on sustained economic expansion, upward movement in real income levels, a strengthening of government capabilities, and a sustained ability of government to co-opt leaders from newly urbanized populations. Without success in these areas, some researchers suggested, the potential for political strife could be far higher.⁴⁹

Later studies touched upon the relationship between urban growth and violence in times of economic stress and offered hints that a causal link might in fact exist. For instance, an examination of Mexican squatter settlements during the 1980s showed that squatters in shantytowns were especially prone to random violence in economically hard times.⁵⁰ Compared with residents of inner city slums, squatters were far from major commercial markets and were therefore disadvantaged in economic competition in the informal sector. Unable to afford income-producing land plots, and lacking basic social and urban services given the reduced outlays of financially strapped governments, these settlements became “slums of despair,” with crime and violence an increasing part of their day to day life.⁵¹

Several studies of the widespread rioting, demonstrations, and protest in Third World cities during the debt crisis of the late 1970s and early 1980s also stressed the importance of the general economic context. These studies attribute the protest to the austerity measures adopted by Third World countries under pressure from the International Monetary Fund (IMF) and the World Bank. The austerity measures eliminated mechanisms for distributing wealth established by governments during years of heavy borrowing in the 1960s and 1970s, and they imposed severe hardships on the urban poor and working classes. The measures were also often perceived as a violation of certain tacit norms of justice and fairness – a moral economy – that had emerged under previous economic arrangements.⁵² Destruction of these arrangements prompted a wave of strikes, riots, and demonstrations throughout the developing world. The timing and intensity of these protests correlated strongly with rural-urban migration.⁵³

Declining interest in the links between urban growth and violence meant that researchers did not pursue evidence pointing to a more complex causal relationship. In recent years, there has been little explicit, critical analysis of the problem.

Possible Interactive Factors

Notwithstanding early research results and the current lack of scholarly interest in the issue, the potential links between urban growth and violence still deserve attention. Analysts have not yet adequately examined the possibility that urban growth can significantly contribute to violence through a synergistic or multiplicative interaction with other factors. Studies that show urban protest and rioting in times of rapid rural-urban migration and economic recession suggest that both factors are needed for the violence to occur. If so, each factor is insufficient to produce a given instance of urban violence, but both are necessary; moreover, it is meaningless to ask about the relative causal contributions or weights of the different factors.⁵⁴

We need to identify, therefore, key factors that might interact with urban growth to produce violence. Scholars have not produced an exhaustive list, but a close look at trends in developing countries reveals several forces that merit attention. When taken as a whole – as a possible “interactive set” – these factors point to an urban future more violent than the past.

First, we have already mentioned the economic crises that afflict many developing states. Some analysts stress the effects on poor countries of the international economic system, especially the effects of structural adjustment policies imposed by international financial institutions such as the IMF and the World Bank.⁵⁵Others propose more complicated causality, with economic problems the result of both international and domestic forces.⁵⁶ International economic recession, entrenched subsidies to domestic vested interests, weak financial and regulatory institutions at the national level, and fiscal and monetary mismanagement by the state all contribute their share. Chronic economic problems often lead to rising grievances within the population, especially among marginalized groups such as urban slum dwellers, and to conflicts within and among elites as subsidies and other opportunities for capturing economic “rents” decline. Protracted economic crisis squeezes tax revenues, and thereby weakens all state institutions, including the bureaucracy, the judiciary, the police, and the military; cutbacks and lower salaries encourage corruption within the civil service. The state becomes evermore unable to meet the demands of competing elites or the grievances of the middle and lower classes, and it cannot confront potential challenger groups with sufficient coercive force. Opportunities for popular protest and rebellion correspondingly increase.

Second, whether economic crisis produces institutional breakdown and violent action depends on the character of civil society, particularly on the society’s stock of “social capital.” This is the accumulation of social organizations, groups, networks, and affiliations based on trust and norms of reciprocity. These organizations can be either formal or informal and can include, for example, churches and mosques, unions, mutual-aid societies, clubs and cooperatives. Such social capital facilitates coordinated action among individuals and groups, and it improves transaction efficiency.⁵⁷ It also provides a degree of cooperation and social solidarity that can buffer people from the harshest effects of economic and state crisis. Unfortunately, in many cities in the developing world, networks of reciprocity are torn by fierce rivalries between communal and ethnic groups, economic classes, and criminal gangs; the social capital available may therefore be insufficient for the staggering demands placed upon it.

A third and related factor is the degree of communalism or ethnic cleavage present in the society. Such cleavages are now less suppressed by the ideological rigidities of the Cold War. Analysts must take note of the history and character of the animosity between rival groups, the degree of their political and economic inequality, and whether there are legal and institutional arrangements that sustain or mitigate this inequality. These conditions will influence not only the likelihood and intensity of conflict but also the state’s capacity to address it.

Growing demands for democratization from mobilized elements of society are a fourth factor that might interact with urban growth to produce violence. Economic crisis and decaying state capacity often boost pressure for democratization, since effective solutions to these problems often entail sacrifices from citizens, and citizens see an expansion of political participation as a reasonable quid pro quo.⁵⁸ All too often, though, the initial response by the state to calls for more democracy is repression. If peaceful channels of political expression are foreclosed, protest can be increasingly violent.

Fifth, the growing strength and reach of organized crime in the developing world compound the stress of urban growth. In many poor societies, the balance of power seems to be shifting from the state and its coercive institutions toward gangs and organized crime. A vigorous global trade in light weapons and plastic explosives means that criminals have easy access to the means of violence. They are often better armed and organized than financially strapped and technologically backward police forces. Moreover, organized crime often penetrates into the heart of the state through blackmail, bribery, and threats of violence. Today, criminal organizations are so well entrenched in some Latin American and South Asian cities that their power clearly exceeds that of local authorities; in Rio de Janeiro and Karachi, the central government has had to use the army to reassert control.

Sixth, as indicated under the preceding point, at the same time that motivations for urban violence have increased in some societies, it is easier for aggrieved groups and criminals to get hold of the means of violence. With the end of the Cold War, some regions are awash in small arms; for example, weapons left over from the Afghan war have saturated Karachi. In addition, many former communist states are eager suppliers of weapons for hard currency. Whatever the source, the result is often an increase in the frequency and destructiveness of urban violence.

The above list of interactive factors does not, by itself, overturn the negative findings of past research. Yet some cities today provide strong prima facie evidence for the claim that various interactive combinations of urban growth and other factors will sharply increase urban violence in the future. Karachi, Pakistan’s largest city and sole industrial port, is a good example of how multiple forces converge to produce explosive results.

Successive waves of migration to Karachi have produced a city of considerable ethnic diversity. After the division of the Indian subcontinent in 1947, the city’s indigenous Sindhi population was overtaken by Urdu-speaking refugees (Muhajirs). In the 1960s, the government’s green-revolution and industrialization policies caused a wave of Pashto speakers (Pathan) to move in from Pakistan’s northern provinces. These changes have marginalized Karachi’s Sindhi population both linguistically and culturally, yet migrant groups are underrepresented within the provincial bureaucracy.⁵⁹

Today, Urdu is the lingua franca of Karachi, and Muhajirs run much of the city’s business and industry. The Pathan make up the majority of the working class and have gained a virtual monopoly over Karachi’s transport sector. Retaining deeply rooted tribal traditions and support systems, they are in effect a separate state within the city. Meanwhile, the Sindhi minority dominate government and educational institutions through a system of quotas.⁶⁰ Rivalries among these groups are common and flow largely from the positions in society that the groups occupy. The presence of contending religious sects further compounds conflict: Shia-Sunni confrontations occur with almost ritualized regularity.

The Pakistani state at the national, provincial, and local levels lacks the capacity and basic institutions needed to accommodate the needs of Karachi’s diverse and quarreling population. Pakistan’s overdeveloped military-bureaucratic oligarchy is rife with corruption and patronage; truly independent and representative political institutions have never been developed at any level of governance; and few institutions are available to ease the transition of migrants to urban life. Local government is characterized by murky lines of authority, few taxing powers, and little accountability.

A high urban growth rate – at about 400,000 people per year – accentuates these problems by intensifying popular demands, underlining state impotence, and further polarizing society. As Karachi’s population rises at 6 percent per annum (far above the national rate of around 3 percent), urban services expand by only 1.2 percent.⁶¹ Government and development authorities cannot provide residents with basic services; the city’s aging infrastructure is overtaxed and does not properly service new communities. Housing is in critically short supply: the government is able to meet only about one-eighth of total annual demand.⁶² Meanwhile, an informal system of illegal occupation and subdividing of state land for sale to low-income families has developed. Managed by middlemen and corrupt government officials, this system often defies state regulations.

Acute shortages of electricity and water are pervasive. Powerful communities in the city are able to insure better service through political pressure and bribes and by building private wells and electrical generators, but those in less fortunate areas are ignored. Karachi’s transportation system, which is largely made up of privately owned minibuses, also stands in disarray. Operators work long hours to repay loans taken for bus purchase. Traffic laws and established transport routes are routinely violated, passengers are mistreated, and accidents are common. The result is acute public resentment, both of the government for its inaction and of an overwhelmingly Pathan core of operators.⁶³

Adding to the problems is the ever present influence of organized crime. Trafficking in narcotics and arms has gained a foothold in many parts of the city, since Karachi is an exit point for the narcotics trade to the rest of the world. Working with Afghan refugees and corrupt government and police officials, Mafia-type syndicates prey on the city’s weaknesses and are adept at exploiting ethnic and religious rivalries to block state challenges to their power.⁶⁴ All the while, the accessibility to inexpensive armaments has increased, especially following the Afghan war, which has magnified the potential for violence among rival groups.

A general climate of insecurity pervades the city. Horizontal polarization of ethnic and religious groups and vertical polarization by economic strata continue unabated. Meanwhile, much-needed investment is driven elsewhere – further eroding the city’s economic base. The climate of insecurity has also crippled Karachi’s educational system. Some colleges have been forced to close, and others now serve as armed strongholds for warring factions; education has become privatized and increasingly segregated along class lines.

With institutionalized channels of protest and action on grievances unavailable, state legitimacy suffers. Popular loyalties and allegiances remain local, and efforts to redress grievances often take the form of ethnic and class-based violence. Frustration stemming from the lack of urban services has reportedly prompted frequent attacks on the offices of the Karachi Electricity Supply Corporation and the Karachi Water and Sewage Board.⁶⁵ Minibus accidents spark ethnic riots (Table 6), and fights between Karachi residents and an underfunded police force are common. The fact that the police are heavily drawn from the northern provinces heightens ethnic tension.⁶⁶

Overall, therefore, violence in Karachi stems from a variety of factors that interact to magnify the impacts each might produce separately. The inability of state institutions to address diverse demands brings to the fore latent ethnic and class tensions. High urban population growth further boosts grievances, highlights the impotence of the state, and reduces state legitimacy. The vacuum left by state weakness strengthens ethnic and class-based loyalties, which further polarizes society. Criminal elements exploit state weakness and societal conflicts to their advantage and import small arms that raise the ability of all contending parties to engage in violence. All the while, the social fabric and economy continue to erode, and state and social capacities to address the plethora of problems further diminish.

Table 6: Karachi – Frequency of and Deaths from Rioting and Violence, January 1986-August 1987

Cause	Number of Incidents	Number of Deaths
Electricity breakdown	1	–
Water shortage	4	1
Transport problem	17	78
Political motivated banditry	80	34
Politically motivated bombing	76	188
Others	102	–
Total	242	403
Source: Akmal Hussain, “The Karachi Riots of December 1986: Crisis of State and Civil Society in Pakistan,” in Veena Das, ed., Mirrors of Violence: Communities, Riots and Survivors in South Asia (New York: Oxford University Press, 1990), p. 189.

Two Hypotheses For Futher Study

Early work on urban growth and violence may not have been misguided so much as too simplistic. It posited links between these variables while paying little attention to other political, economic, and social factors. In light of our discussion to this point, we propose two additional hypotheses about the links between urban growth and violence that deserve further study.

First, as noted above, international economic pressures (such as structural adjustment programs) combined with misguided domestic economic policies can undermine a state’s ability to cope with the demands of a growing urban populace. With economic crisis, state capacity to manipulate the opportunities available to potential challengers may decrease; the social balance of power may shift in the challengers’ favor. Worsening economic conditions and reduced state capacity might also combine to increase the intensity of grievances among the urban populace. The effects of economic problems would likely be stronger if accompanied by rapidly expanding cities, since this growth would further boost demands on the state and accentuate grievances. (Figure 4).

Although rural-urban migration might decline as economic problems reduce opportunities in cities, much would depend on whether rural areas were also affected by recession and on the rural impacts of rapid population growth and resource scarcity. Africa shows that rapid rural-urban migration does occur during economic downturns.

A second hypothesis focuses more on the urban born rather than on rural-urban migrants. Early research focused on migrants. Some experts argued that migrants need a “socialization period” to be radicalized by urban life; they may need to leave behind earlier – often less favorable – rural experiences. Only then will they experience deprivation and frustration.

This idea can be extended. In the future, as societies become increasingly urbanized and rural ways of life decline, there may be ever fewer opportunities to draw favorable comparisons between urban and rural lifestyles. In fact, a very large majority of people will be born in cities and will have had no rural experience. Urban life will increasingly represent the only arena of comparison for the masses. In the context of economic stagnation or recession, relative differences between rich and poor in the city, and between different ethnic groups and classes, will become evermore salient in people’s minds. In these circumstances, feelings of relative deprivation are likely to rise (Figure 5).

For example, the street crime in many Latin American cities suggests a growing tendency toward violence among better-educated, urban-born males with few avenues for economic advancement.⁶⁷ Expectations outstrip opportunities. Urban growth plays a role, not by creating angry communities of recent migrants from rural areas, but rather by creating a glut of young, urban-born job and status seekers who cannot be satisfied without fast economic expansion.

Criminal violence, however, does not necessarily translate into the organized political violence that students of political conflict emphasize. But it bears repeating that more frequent acts of individual or gang violence may indicate a general breakdown of societal norms and state legitimacy that could eventually translate into broadly based movements against the state.

Conclusion

There are no simple links between urban growth and violence: urban violence is influenced by a wide array of factors that interact in complex ways. Researchers must therefore undertake detailed case studies if they are to understand the links among these variables.

Nevertheless, cities will probably be the locus of much of the future conflict and violence in the developing world, particularly in the context of economic recession and readjustment, declining state capacity, and growing demands for democratization. When urban areas and states face converging economic and social pressures, power and privilege centered in the city will be challenged. Although rural-urban migration will continue and will, in many cases, magnify the social and economic problems of cities, it seems likely that the participants in urban violence will be urban born.

The Environment, Population and Security papers are maintained by the Peace & Conflict Studies Program at the University of Toronto.

Appendix:
The World’s Urban Populations – 1950 to 2010

Data calculated from the World Resources 1994-95 Data Base Diskettes, World Resources Institute, Washington D.C., 1994. For each 15-year time period, the “Average Annual % Increase” is calculated as the mean of three means: the mean urban population growth rates for the three consecutive 5-year periods that make up the 15 years.

Countries & Regions	% of Total 1950	% of Total 1995	% of Total 2010	Average Annual % Increase 1950 to 1965	Average Annual % Increase 1965 to 1980	Average Annual % Increase 1980 to 1995	Average Annual % Increase 1995 to 2010
AFRICA	14.5	34.7	53.9	4.9	4.8	4.7	4.4
ASIA	16.4	34.0	50.1	4.1	3.4	3.7	3.1
CENTRAL AMERICA	39.7	68.3	82.3	4.8	4.2	3.2	2.4
EUROPE	56.2	75.0	82.2	1.7	1.2	0.7	0.7
NORTH AMERICA	63.9	76.4	85.0	2.8	1.8	1.7	1.5
OCEANIA	61.4	70.9	78.8	2.9	2.0	1.6	1.6
SOUTH AMERICA	43.2	78.0	90.0	4.6	3.7	2.8	1.8
WORLD	29.3	45.2	59.2	3.2	2.7	2.7	2.5
Afghanistan	5.8	20.0	28.2	5.5	5.5	4.3	4.9
Albania	20.3	37.3	45.7	5.7	2.9	2.3	2.5
Algeria	22.3	55.8	66.4	5.7	4.2	4.6	3.6
Angola	7.6	32.2	44.2	5.2	5.9	6.2	5.3
Antigua and Barbuda	45.7	33.8	46.1	0.2	-0.8	1.4	2.8
Argentina	65.3	87.5	90.6	2.8	2.2	1.6	1.3
Australia	75.1	85.2	87.0	2.8	1.9	1.4	1.4
Austria	49.1	60.6	68.7	0.6	0.8	1.0	1.1
Bahrain	63.8	84.3	87.6	5.2	4.0	3.7	2.5
Bangladesh	4.2	19.5	30.3	5.3	7.2	6.4	5.2
Barbados	33.6	47.9	58.5	1.2	1.1	1.5	1.9
Belgium	91.5	96.7	97.4	0.7	0.4	0.2	0.1
Belize	56.7	52.6	62.1	2.7	1.7	2.9	2.6
Benin	4.3	41.8	53.1	9.4	8.6	5.0	4.6
Bhutan	2.0	6.4	11.4	3.9	4.6	5.9	6.4
Bolivia	37.8	54.4	64.3	2.7	3.3	3.9	3.3
Botswana	0.3	30.8	46.9	26.1	13.6	8.2	5.5
Brazil	36.0	78.7	85.3	5.5	4.3	3.0	1.7
Brunei	27.1	57.6	64.3	9.9	5.1	2.4	2.3
Bulgaria	25.6	70.7	78.0	4.7	2.4	1.0	0.6
Burkina Faso	3.8	19.5	33.5	4.4	6.3	8.7	6.4
Burundi	1.7	6.1	9.8	4.0	6.4	5.6	6.0
Cambodia	10.2	12.9	19.7	2.6	0.4	4.3	5.1
Cameroon	9.8	44.9	57.4	5.8	7.2	5.4	4.5
Canada	60.8	78.1	82.1	3.6	1.6	1.4	1.5
Cape Verde	8.2	32.0	43.5	8.3	3.6	4.7	4.5
Central African Republic	16.0	50.7	61.5	5.3	4.8	4.6	3.7
Chad	3.8	37.0	49.7	7.9	7.9	6.5	4.6
Chile	58.4	85.9	89.2	3.7	2.6	2.0	1.5
China	11.0	30.2	43.0	5.2	2.8	4.4	3.2
Colombia	37.1	72.7	79.4	5.5	3.6	2.7	1.9
Comoros	3.5	30.8	41.8	13.3	5.8	5.8	5.5
Congo	30.9	43.4	54.0	2.6	3.6	4.4	4.3
Costa Rica	33.5	49.7	59.4	4.6	3.8	3.7	3.1
Cote d’Ivoire	13.1	43.6	54.1	7.6	7.0	5.5	4.9
Cuba	49.4	76.0	81.8	3.0	2.6	1.6	1.1
Cyprus	29.8	56.3	65.9	2.7	1.9	2.4	1.8
Czechoslovakia (former)	37.4	80.6	87.0	3.0	2.4	1.4	1.0
Denmark	68.0	85.5	88.0	1.5	1.0	0.2	0.3
Djibouti	41.7	82.8	86.7	7.0	8.7	4.3	3.2
Dominican Rep	23.8	64.6	73.5	6.1	5.3	3.8	2.4
Ecuador	28.2	60.6	70.9	5.0	4.8	4.2	2.8
Egypt	31.9	44.8	51.8	4.1	2.8	2.6	2.9
El Salvador	36.5	46.7	56.3	3.4	3.2	2.5	3.3
Equatorial Guinea	15.9	30.5	40.2	4.3	-0.7	5.0	4.4
Estonia	49.4	73.1	78.0	2.5	1.7	0.7	0.7
Ethiopia	4.6	13.4	19.5	5.7	4.7	4.7	5.6
Fiji	24.2	40.7	48.7	5.2	3.0	1.8	2.2
Finland	32.0	60.3	65.7	3.0	2.3	0.4	0.8
France	56.2	72.8	76.3	2.2	1.2	0.4	0.6
French Guiana	52.0	76.3	82.1	4.8	4.4	3.9	2.8
Gabon	11.3	50.0	60.7	5.1	7.1	5.9	4.0
Gambia	10.5	25.5	36.7	4.1	5.3	5.3	4.9
Germany	71.9	86.4	89.5	1.2	0.6	0.5	0.5
Ghana	14.5	36.3	46.6	7.1	3.5	4.5	4.6
Greece	37.3	65.0	72.7	2.5	2.1	1.2	0.8
Greenland	78.3	79.3	84.1	3.4	1.6	1.3	0.9
Guatemala	29.5	41.5	51.3	3.9	3.5	3.7	4.2
Guinea	5.5	29.6	41.6	7.4	5.3	5.9	5.3
Guinea-Bissau	10.1	22.2	32.7	2.8	4.0	4.1	4.8
Guyana	28.1	35.4	46.6	3.0	1.3	1.9	3.0
Haiti	12.2	31.6	42.3	4.2	3.9	4.0	4.1
Honduras	17.6	47.7	59.0	6.2	5.6	5.3	4.0
Hong Kong	82.5	95.0	96.6	4.5	2.4	1.3	0.5
Hungary	39.3	67.7	75.8	1.6	1.8	1.0	0.8
Iceland	74.1	91.8	93.9	2.7	1.6	1.3	1.1
India	17.3	26.8	33.8	2.9	3.6	3.1	3.2
Indonesia	12.4	32.5	44.5	3.8	4.8	4.6	3.4
Iran	27.0	60.4	69.7	5.1	5.1	5.0	4.0
Iraq	35.1	74.6	81.1	5.6	5.1	4.2	3.5
Ireland	41.1	58.4	65.0	1.0	1.9	0.5	0.8
Israel	64.6	92.7	94.5	6.3	3.5	3.0	1.4
Italy	54.3	70.5	76.6	1.5	1.0	0.6	0.6
Jamaica	26.7	55.4	64.8	3.8	2.8	2.3	2.2
Japan	50.3	77.9	81.6	3.1	1.9	0.6	0.5
Jordan	34.7	71.5	79.2	5.2	4.6	4.6	3.8
Kenya	5.6	27.7	39.7	6.5	8.4	7.4	5.6
Kiribati	9.4	39.2	50.0	7.6	4.7	3.8	3.6
North Korea	31.0	61.3	67.8	4.3	4.1	2.3	2.0
South Korea	21.4	77.6	86.3	5.4	5.8	3.1	1.2
Kuwait	59.2	97.0	98.2	10.3	8.1	1.6	2.2
Laos	7.2	21.7	32.6	3.4	5.5	6.3	5.3
Latvia	51.4	72.8	78.2	2.0	1.6	0.7	0.6
Lebanon	22.7	87.2	92.1	8.1	4.0	2.0	1.8
Lesotho	1.0	23.1	35.3	15.9	7.6	6.6	5.3
Liberia	13.0	50.6	63.3	6.4	6.4	5.9	4.7
Libya	18.6	86.0	90.7	7.0	10.9	5.3	3.6
Lithuania	30.9	72.1	79.5	3.5	3.0	1.7	1.0
Luxembourg	59.1	86.0	89.2	1.3	2.0	1.0	0.8
Madagascar	7.8	27.1	38.6	5.7	5.5	6.1	5.6
Malawi	3.5	13.5	21.0	5.0	7.7	7.0	5.7
Malaysia	20.4	47.2	58.4	4.8	4.6	4.7	3.2
Mali	8.5	27.0	38.1	5.1	5.1	5.8	5.4
Malta	61.2	88.6	91.2	1.1	1.3	1.2	0.9
Mauritania	2.3	53.8	65.3	12.5	10.8	7.0	4.0
Mauritius	28.8	40.7	46.7	4.5	2.5	0.9	1.8
Mexico	42.7	75.3	81.6	4.9	4.3	3.1	2.1
Mongolia	18.9	60.9	69.4	8.0	4.3	3.9	3.2
Morocco	26.2	48.4	56.8	4.2	4.3	3.7	3.2
Mozambique	2.4	34.3	50.5	7.5	10.5	8.8	5.6
Myanmar	16.2	26.2	35.4	3.8	3.2	2.9	4.0
Namibia	9.4	30.9	42.1	6.3	5.1	5.4	5.1
Nepal	2.3	13.7	23.2	4.9	7.3	8.0	5.9
Netherlands	82.7	88.9	90.6	1.5	1.1	0.7	0.7
New Zealand	72.5	84.3	86.8	2.7	1.5	1.0	1.0
Nicaragua	34.9	62.9	71.3	4.7	4.8	4.3	3.7
Niger	4.9	23.1	34.4	5.7	7.8	7.4	5.9
Nigeria	10.1	39.3	51.1	6.5	6.7	5.9	4.8
Norway	50.1	77.0	82.3	1.8	1.9	1.0	0.9
Oman	2.4	13.2	21.7	6.5	8.3	8.3	7.0
Pakistan	17.5	34.7	45.4	4.6	4.1	4.7	4.4
Panama	35.7	54.9	62.8	4.2	3.4	2.7	2.4
Papua New Guinea	0.7	17.8	26.7	17.4	8.9	4.6	5.0
Paraguay	34.6	50.7	60.8	3.3	4.0	4.4	3.6
Peru	35.5	72.2	78.6	5.4	4.3	2.9	2.3
Philippines	27.1	45.7	55.7	4.0	4.0	3.8	3.0
Poland	38.7	63.9	70.9	3.3	1.8	1.2	1.2
Portugal	19.2	36.4	47.2	2.0	1.8	1.5	1.9
Romania	25.5	56.2	64.7	3.7	2.7	1.3	1.4
Rwanda	1.8	6.1	8.9	6.2	7.1	5.2	5.9
Saudi Arabia	15.9	80.2	84.7	9.5	8.5	5.5	3.5
Senegal	30.5	42.3	51.7	3.1	3.6	4.0	4.0
Seychelles	26.5	64.9	74.1	2.3	5.5	3.8	1.5
Sierra Leone	9.2	36.2	48.1	5.2	5.4	5.3	4.5
Singapore	100.0	100.0	100.0	4.0	1.7	1.1	0.7
Solomon Islands	7.8	17.2	26.7	3.8	5.1	7.1	6.1
Somalia	12.7	25.7	33.9	5.1	4.1	4.0	5.0
South Africa	43.1	50.8	59.5	3.2	2.9	2.9	3.2
Spain	51.9	80.7	85.5	2.1	2.2	1.0	0.6
Sri Lanka	14.4	22.4	30.7	4.6	2.4	1.8	3.3
Sudan	6.3	24.6	34.4	7.2	5.8	4.6	5.0
Suriname	47.0	50.3	61.0	2.8	0.3	2.7	2.6
Swaziland	1.5	31.2	45.4	13.5	10.2	6.8	5.2
Sweden	65.8	84.7	87.7	1.7	1.0	0.5	0.6
Switzerland	44.3	64.0	71.4	2.6	1.1	1.4	1.2
Syria	30.6	52.4	60.7	4.8	4.5	4.5	4.4
Tanzania	3.8	24.4	36.3	6.1	10.6	7.1	5.9
Thailand	10.5	25.4	36.6	4.5	4.9	4.2	3.4
Togo	7.2	30.8	40.6	5.1	8.2	5.3	4.9
Tonga	12.2	41.4	56.3	7.3	2.8	4.2	2.8
Trinidad and Tobago	64.0	66.6	73.8	2.2	1.2	1.6	1.7
Tunisia	31.2	59.0	67.9	3.5	3.9	3.3	2.5
Turkey	21.3	68.8	82.4	6.0	4.4	5.3	2.8
U.S.S.R. (former)	41.5	68.1	74.5	3.3	2.0	1.3	1.2
Uganda	3.1	12.5	18.8	8.8	5.6	5.6	5.6
United Arab Emirates	24.3	84.0	88.7	13.3	17.0	4.8	2.2
United Kingdom	84.2	89.5	91.4	0.7	0.4	0.3	0.3
United States	64.2	76.2	80.4	2.3	1.2	1.2	1.1
Uruguay	78.0	90.3	93.0	1.5	0.9	1.0	0.7
Vanuatu	5.8	19.5	25.8	7.0	6.5	3.2	4.5
Venezuela	53.2	92.9	96.3	5.9	4.7	3.1	1.9
Vietnam	11.6	20.8	27.4	4.0	3.4	2.8	3.8
Yemen	5.8	33.6	47.4	6.8	6.9	7.3	5.6
Yugoslavia (former)	22.0	60.7	70.7	3.6	3.4	2.4	1.4
Zaire	19.1	29.1	37.9	4.6	3.6	3.6	4.9
Zambia	8.9	43.1	49.7	9.7	6.8	3.9	3.7
Zimbabwe	10.7	32.1	43.9	5.7	6.4	5.9	4.7

Endnotes

*The authors thank Alan Gilbert, Josef Guglar, Joan Nelson, and John Walton for their comments on an earlier version of this paper.